Skip to main content

Table 3 Detailed comparison of the US 2022 physician workforce report data with AI-generated images

From: Representation of intensivists’ race/ethnicity, sex, and age by artificial intelligence: a cross-sectional study of two text-to-image models

 

US 2022 physician workforce report

Midjourney

P value

ChatGPT DALL-E 2

P value

Female sex (n, %)

 Overall

6′972/21′678 (32.2%)

333/700 (47.6%)

< 0.001

121/700 (17.3%)

< 0.001

 Pediatric intensivists

1′556/3′099 (50.2%)

57/100 (57.0%)

0.181

48/100 (48.0%)

0.664

 Surgical intensivists

1′064/2′980 (35.7%)

58/100 (58.0%)

< 0.001

1/100 (1.0%)

< 0.001

 Medical intensivists

4′352/15′599 (27.9%)

90/300 (30.0%)

0.422

58/300 (19.3%)

0.001

White (n, %)*

 Overall

11′950/21′678 (55.1%)

426/700 (60.9%)

0.003

427/700 (61.0%)

0.002

 Pediatric intensivists

1′909/3′099 (61.6%)

61/100 (61.0%)

0.903

427/700 (61.0%)

< 0.001

 Surgical intensivists

1′961/2′980 (65.8%)

70/100 (70.0%)

0.384

52/100 (52.0%)

0.004

 Medical intensivists

8′080/15′599 (51.8%)

172/300 (57.3%)

0.057

186/300 (62.0%)

< 0.001

Age < 40 years (n, %)

 Overall

5′181/21′678 (23.9%)

345/700 (49.3%)

< 0.001

373/700 (53.3%)

< 0.001

 Pediatric intensivists

725/3′099 (23.4%)

83/100 (83.0%)

< 0.001

67/100 (67.0%)

< 0.001

 Surgical intensivists

790/2′980 (26.5%)

43/100 (43.0%)

< 0.001

54/100 (54.0%)

< 0.001

 Medical intensivists

3′666/15′599 (23.5%)

109/300 (36.3%)

< 0.001

152/300 (50.7%)

< 0.001

  1. *Individuals classified as Black, Asian, Hispanic/Latino, Multiracial and Other were categorized as a composite “non-White” group. P values reflect the comparison between the US 2022 physician workforce report data and the respective AI text-to-image models (Midjourney and ChatGPT DALL-E 2) across the different demographic categories